Published: May 18, 2011 r2011 American Chemical Society 5323 dx.doi.org/10.1021/es2007462 | Environ. Sci. Technol. 2011, 45, 5323–5331 ARTICLE pubs.acs.org/est Identification of Flame Retardants in Polyurethane Foam Collected from Baby Products Heather M. Stapleton,* ,† Susan Klosterhaus, ‡ Alex Keller, † P. Lee Ferguson, † Saskia van Bergen, § Ellen Cooper, † Thomas F. Webster, || and Arlene Blum ^ † Nicholas School of the Environment, Duke University, Durham, North Carolina, United States ‡ San Francisco Estuary Institute, Oakland, California, United States § East Bay Municipal Utility District, Oakland, California, United States ) Department of Environmental Health, Boston University School of Public Health, Boston, Massachusetts, United States ^ Department of Chemistry, University of California, and Green Science Policy Institute, Berkeley, California, United States b S Supporting Information ’ INTRODUCTION Prior to 2004, PentaBDE was one of the most common flame retardant mixtures added to polyurethane foam in furniture and other consumer products, particularly in the US. Because of concerns regarding the persistence, bioaccumulation, and poten- tial toxicity of the polybrominated diphenyl ethers (PBDEs) present in this commercial mixture, California passed legislation banning its use in 2003. Eight other states and the European Union (EU) followed with similar bans and the sole U.S. manufacturer, Great Lakes Chemical (now Chemtura), voluntarily phased out production in 2004. 1,2 Alternative chemical flame retardants have since been used and identified as PentaBDE replacements in polyurethane foam. 3,4 However, basic information on these alter- native flame retardants, such as chemical identity, specific pro- duct applications, and volumes used, are typically not available, Received: March 4, 2011 Accepted: April 29, 2011 Revised: April 26, 2011 ABSTRACT: With the phase-out of PentaBDE in 2004, alternative flame retardants are being used in polyurethane foam to meet flammability standards. However, insufficient information is available on the identity of the flame retardants currently in use. Baby products containing polyurethane foam must meet California state furniture flammability standards, which likely affects the use of flame retardants in baby products throughout the U.S. However, it is unclear which products contain flame retardants and at what concentrations. In this study we surveyed baby products containing polyurethane foam to investigate how often flame retardants were used in these products. Information on when the products were purchased and whether they contained a label indicating that the product meets requirements for a California flammability standard were recorded. When possible, we identified the flame retardants being used and their concentrations in the foam. Foam samples collected from 101 commonly used baby products were analyzed. Eighty samples contained an identifiable flame retardant additive, and all but one of these was either chlorinated or brominated. The most common flame retardant detected was tris(1,3-dichloroisopropyl) phosphate (TDCPP; detection frequency 36%), followed by components typically found in the Firemaster550 commercial mixture (detection frequency 17%). Five samples contained PBDE congeners commonly associated with PentaBDE, suggesting products with PentaBDE are still in-use. Two chlorinated organophosphate flame retardants (OPFRs) not previously documented in the environment were also identified, one of which is commercially sold as V6 (detection frequency 15%) and contains tris(2-chloroethyl) phosphate (TCEP) as an impurity. As an addition to this study, we used a portable X-ray fluorescence (XRF) analyzer to estimate the bromine and chlorine content of the foam and investigate whether XRF is a useful method for predicting the presence of halogenated flame retardant additives in these products. A significant correlation was observed for bromine; however, there was no significant relationship observed for chlorine. To the authors knowledge, this is the first study to report on flame retardants in baby products. In addition, we have identified two chlorinated OPFRs not previously documented in the environment or in consumer products. Based on exposure estimates conducted by the Consumer Product Safety Commission (CPSC), we predict that infants may receive greater exposure to TDCPP from these products compared to the average child or adult from upholstered furniture, all of which are higher than acceptable daily intake levels of TDCPP set by the CPSC. Future studies are therefore warranted to specifically measure infants exposure to these flame retardants from intimate contact with these products and to determine if there are any associated health concerns.
Transcript
Published: May 18, 2011
r 2011 American Chemical Society 5323 dx.doi.org/10.1021/es2007462 | Environ. Sci. Technol. 2011, 45, 5323–5331
ARTICLE
pubs.acs.org/est
Identification of Flame Retardants in Polyurethane Foam Collectedfrom Baby ProductsHeather M. Stapleton,*,† Susan Klosterhaus,‡ Alex Keller,† P. Lee Ferguson,† Saskia van Bergen,§
Ellen Cooper,† Thomas F. Webster,|| and Arlene Blum^
†Nicholas School of the Environment, Duke University, Durham, North Carolina, United States‡San Francisco Estuary Institute, Oakland, California, United States§East Bay Municipal Utility District, Oakland, California, United States
)Department of Environmental Health, Boston University School of Public Health, Boston, Massachusetts, United States^Department of Chemistry, University of California, and Green Science Policy Institute, Berkeley, California, United States
bS Supporting Information
’ INTRODUCTION
Prior to 2004, PentaBDE was one of the most common flameretardant mixtures added to polyurethane foam in furniture andother consumer products, particularly in the US. Because ofconcerns regarding the persistence, bioaccumulation, and poten-tial toxicity of the polybrominated diphenyl ethers (PBDEs)present in this commercial mixture, California passed legislationbanning its use in 2003. Eight other states and the EuropeanUnion(EU) followed with similar bans and the sole U.S. manufacturer,Great Lakes Chemical (now Chemtura), voluntarily phased out
production in 2004.1,2 Alternative chemical flame retardants havesince been used and identified as PentaBDE replacements inpolyurethane foam.3,4 However, basic information on these alter-native flame retardants, such as chemical identity, specific pro-duct applications, and volumes used, are typically not available,
Received: March 4, 2011Accepted: April 29, 2011Revised: April 26, 2011
ABSTRACT:With the phase-out of PentaBDE in 2004, alternative flame retardants are being used inpolyurethane foam to meet flammability standards. However, insufficient information is available onthe identity of the flame retardants currently in use. Baby products containing polyurethane foammust meet California state furniture flammability standards, which likely affects the use of flameretardants in baby products throughout the U.S. However, it is unclear which products contain flameretardants and at what concentrations. In this study we surveyed baby products containingpolyurethane foam to investigate how often flame retardants were used in these products.Information on when the products were purchased and whether they contained a label indicatingthat the product meets requirements for a California flammability standard were recorded. Whenpossible, we identified the flame retardants being used and their concentrations in the foam. Foamsamples collected from 101 commonly used baby products were analyzed. Eighty samples containedan identifiable flame retardant additive, and all but one of these was either chlorinated or brominated.The most common flame retardant detected was tris(1,3-dichloroisopropyl) phosphate (TDCPP;detection frequency 36%), followed by components typically found in the Firemaster550 commercial mixture (detection frequency17%). Five samples contained PBDE congeners commonly associated with PentaBDE, suggesting products with PentaBDE are stillin-use. Two chlorinated organophosphate flame retardants (OPFRs) not previously documented in the environment were alsoidentified, one of which is commercially sold as V6 (detection frequency 15%) and contains tris(2-chloroethyl) phosphate (TCEP)as an impurity. As an addition to this study, we used a portable X-ray fluorescence (XRF) analyzer to estimate the bromine andchlorine content of the foam and investigate whether XRF is a useful method for predicting the presence of halogenated flameretardant additives in these products. A significant correlation was observed for bromine; however, there was no significantrelationship observed for chlorine. To the authors knowledge, this is the first study to report on flame retardants in baby products. Inaddition, we have identified two chlorinatedOPFRs not previously documented in the environment or in consumer products. Basedon exposure estimates conducted by the Consumer Product Safety Commission (CPSC), we predict that infants may receive greaterexposure to TDCPP from these products compared to the average child or adult from upholstered furniture, all of which are higherthan acceptable daily intake levels of TDCPP set by the CPSC. Future studies are therefore warranted to specifically measure infantsexposure to these flame retardants from intimate contact with these products and to determine if there are any associated healthconcerns.
significantly restricting human and environmental health evalua-tions.Many of the chemical ingredients inflame retardantmixturesare proprietary and are not disclosed by the chemical manufac-turers, even to manufacturers using these chemicals in their finalend products (e.g., furniture).
The flammability standard primarily driving the use of flameretardant chemicals in polyurethane foam in the US is TechnicalBulletin 117 (TB117), promulgated by the California Bureau ofElectronic and Appliance Repair, Home Furnishings and Ther-mal Insulation. TB117 requires that polyurethane foam inupholstered furniture sold in the State of California withstandexposure to a small open flame for 12 s.5 Though the standarddoes not specifically require the addition of flame retardantchemicals to the foam, polyurethane foam manufacturers typi-cally use chemical additives as an efficient method for meetingthe TB117 performance criteria.6 Throughout the 1980s and1990s, PentaBDE was used often in the US to comply withTB117. Numerous studies have since documented widespreadcontamination of the PBDE congeners found in the PentaBDEmixture in both humans and wildlife.7,8 PBDEs have also recentlybeen identified in children’s toys.9 Despite the fact that com-pliance with TB117 is only required for residential upholsteredfurniture sold in the State of California, a significant fraction ofproducts sold elsewhere in the US also complies with TB117 andtherefore also contains flame retardant additives.
It is less well-known that some baby products are consideredjuvenile furniture and that the polyurethane foam used in babyproducts must also comply with TB117. However, the extent ofbaby product compliance with TB117 and whether or not thetypes of chemicals added to the polyurethane foam are similar tothose in nonjuvenile furniture is unknown. Flame retardantadditives can escape from products over time, accumulate indust, and are a primary route of exposure to humans.10�13
Exposure to children is a particular concern due to their frequenthand tomouth behavior and higher contact with floors. Exposureto chemical additives in baby products is of even greater concernfor infants, who are in intimate contact with these products forlong periods of time, at very critical stages of their development.Knowledge of the types of chemicals in use and the products theyare used in are essential first steps for evaluating the potential forhuman exposure and subsequent health effects. Structural iden-tities are also needed to track the fate and transport of thesechemicals in the environment.
The objective of this study was to survey a large number ofbaby products that contain polyurethane foam to investigatewhether flame retardant chemicals were present and to deter-mine the concentrations in the foam, in order to understandwhether they may be a significant source of exposure, particularlyto infants. To do this we analyzed foam samples from babyproducts purchased in the US, primarily targeting the mostcommonly used products that contain polyurethane foam. Asecondary objective was to determine whether portable X-rayfluorescence (XRF) is a useful method for predicting thepresence of bromine or chlorinated flame retardant additives inthese products. In a previous study, XRF-measured bromine washighly correlated with gas chromatography�mass spectrometry(GC/MS)-measured bromine in a limited number of pieces offurniture foam and plastics from electronics.12 However, Allenet al. focused on estimating PBDE content, and it is not knownwhether XRF is a useful indicator of the presence of otherbrominated and chlorinated flame retardants. Portable XRFhas potential for use as a less expensive screening tool for
researchers studying potential sources of flame retardant chemi-cals as well as concerned members of the public interested inavoiding products containing flame retardant chemicals. Datagenerated from this study will be useful for informing generalconsumers and scientists about specific flame retardants in use tobetter understand their fate, exposure, and potential healtheffects from using these chemicals in consumer products.
’MATERIALS AND METHODS
Materials. Internal standards were purchased from Chiron(Trondheim, Norway) and Wellington Laboratories (Guelph,Ontario). PBDE calibration standards were purchased fromAccuStandard (New Haven, CT); 2-ethylhexyl-2,3,4,5-tetrabromo-benzoate (TBB) and bis(2-ethylhexyl)-2,3,4,5-tetrabromophthalate(TBPH) were purchased from Wellington Laboratories. Tris(2-chloroethyl) phosphate (TCEP), tris(1-chloro-2-propyl) phosphate(TCPP), and tris(1,3-dichloroisopropyl) phosphate (TDCPP) werepurchased from Sigma-Aldrich (St. Louis, MI), Pfaltz & Bauer(Waterbury, CT), and ChemService (West Chester, PA), respec-tively. All solvents used throughout this study were HPLC grade.Sample Collection. Foam samples were solicited from volun-
teers via email distributions to colleagues and listservs basedprimarily in the United States. Requests were made for samplesof polyurethane foam from baby products, with specific requestsfor samples of car seats, strollers, changing table pads, nursingpillows, portable crib mattresses, and infant sleep positioners.Individuals interested in participating in our study were asked tocut out a small piece of the foam (approximately 2 cm � 2 cm),wrap the foam in aluminum foil, and enclose it in a resealableplastic bag. Participants were also asked to complete a briefsurvey to collect information on the type of product, year ofpurchase, manufacturer, and whether the product possessed alabel indicating that it met the criteria for TB117 or TechnicalBulletins 116 (TB 116) or 603 (TB603). These latter twoCalifornia flammability standards regulate flammability in up-holstered furniture and mattresses, respectively. The sampleswere logged into a database and then split into two pieces, one forchemical analysis by mass spectrometry and one for elementalanalysis using a portable XRF analyzer. Each analysis wasconducted blind.Sample Analysis by Mass Spectrometry. All foam samples
were first screened for flame retardant additives. Briefly, smallpieces of foam (approximately 0.05 g) were sonicated with 1 mLof dichloromethane (DCM) in a test tube for 15 min. The DCMextract was syringe-filtered to remove particles and then trans-ferred to an autosampler vial for analysis by GC/MS. All extractswere analyzed in full scan mode using both electron ionization(GC/EI-MS) and electron capture negative chemical ionization(GC/ECNI-MS). Pressurized temperature vaporization injec-tion was employed in the GC. GC/MS method details can befound in ref 3. All significant peaks observed in the total ionchromatograms were compared to a mass spectral database(NIST, 2005) and to authentic standards when available.If a flame retardant chemical was detected during the initial
screening, a second analysis of the foam sample, using a separatepiece of the foam, was conducted for quantitation using acceler-ated solvent extraction. Our methods for extracting and measur-ing flame retardants in foam are reported in Stapleton et al.3 Afive point calibration curve was established for all analytes withconcentrations ranging from 20 ng/mL to 2 μg/mL. PBDEswere quantified by GC/ECNI-MS by monitoring bromide ions
(m/z 79 and 81), and TBB and TBPH were monitored bymolecular fragments m/z 357/471 and 463/515, respectively.TCEP, TCPP, and TDCPP were quantified by GC/EI-MS bymonitoring m/z 249/251, 277/201, and 381/383, respectively.Because GC/MS analysis of some foam samples suggested the
presence of additional flame retardants that may have beenthermally labile (decomposing partially in the injection port oftheGC) or nonvolatile, all sample extracts were further analyzed byHPLC-high resolution mass spectrometry to determine if addi-tional relevant compounds were present, which were not detectedby GC/MS. HPLC-high resolution mass spectrometry (HPLC/HRMS) analyses were conducted using a LTQ-Orbitrap Velostandem mass spectrometer (ThermoFisher Scientific, Bremen,Germany) with a Thermo Fisher Scientific Accela series UPLCsystem. Sample extracts (25μL) were separated on aHypersil Gold50 � 2.1-mm C18 column with 1.9 μm particles (ThermoFisherScientific) using a flow rate of 0.4 mL/min and a linear gradientfrom 25 to 95%methanol/water in 9min, followed by a 1-min holdat 95% methanol before returning to initial conditions for 2 min.Sample extracts were analyzed using both positive polarity electro-spray ionization (ESI) and atmospheric pressure chemical ioniza-tion (APCI) modes. Prior to analysis, mass calibration wasperformed daily by direct infusion of a calibrationmixture preparedaccording to the instrument manufacturer’s instructions. Massspectral acquisition was programmed into five scan events runningconcurrently throughout the chromatographic separation. The firstscan event was programmed to acquire full-scan (250�2000m/z),high-resolution (R = 60,000) orbitrap MS data with external masscalibration (<2 ppm accuracy). The subsequent four scan eventswere low-resolution data-dependent MS/MS analyses in the LTQion trap analyzer, triggered by the four most intense ions selectedfrom the previous high-resolution orbitrap MS spectrum.XRF Analysis. A portable XRF analyzer (Olympus Innov-X
Systems, Delta model) was used to estimate the elemental com-position of the foam samples. Bromine and chlorine concentrationestimates were obtained using RoHS/WEEE and soil mode,respectively. RoHS/WEEE mode was the only mode availablefor bromine analysis. For chlorine, testing conducted a priori onfoam samples indicated soil mode provided much lower detectionlimits compared to RoHS/WEEE mode. This was supported bythe analysis of the foam samples using RoHS/WEEEmode in thisstudy, which resulted in several nondetect values for chlorinecompared to the use of soil mode. For each sample, three 30 s testswere conducted in each mode sequentially without moving thesample. The average value was used for comparison to GC/MSmeasurements. Though a test stand was not available for use, carewas taken to ensure that the foam sample was flush with theanalyzer window during each test. The original factory instrumentcalibration settings were used. Plastic pellet reference materials(European reference materials EC680K and EC681K) andfurniture foam samples from a previous study3 were analyzedprior to any testing each day and after every 150�200 tests(or ∼25 samples) to ensure there were no substantial changesin instrument performance during testing. Because authenticstandards for polyurethane foam containing bromine andchlorine were not available, XRF data should be consideredsemiquantitative only.
’RESULTS AND DISCUSSION
Identification of Flame Retardants in Foam. A total of 101polyurethane foam samples from baby products were donated forT
use in this study. Most samples were collected from productscurrently in use. However, 14 of the products were purchasednew in 2010 specifically for this study. Samples were donatedfrom participants residing in 13 US states, although one samplewas submitted from Vancouver, Canada. A summary of thenumber and types of products included in this study is shownin Table 1. Most samples were from car seats (n = 21), changingtable pads (n = 16), infant sleep positioners (n = 15), portablecrib mattresses (n = 13), and nursing pillows (n = 11). A fewadditional samples were collected from high chairs, nurseryrocking chairs/gliders, baby walkers, baby carriers, and miscella-neous bathroom items.The chemical structures for the most commonly detected
flame retardants (non-PBDEs) in the baby product foam samplesare presented in Figure 1. Table 1 provides an overview of theflame retardants detected in the baby product foam in concen-trations greater than 1 mg/g. A threshold value of 1 mg/g wasused because while flame retardants are typically added topolyurethane foam at percent levels, some foam samples maycontain flame retardant impurities due to changes in flameretardant applications from batch to batch during foam produc-tion (personal communication from foam manufacturer whowishes to be anonymous). The most common flame retardantdetected was tris(1,3-dichloroisopropyl) phosphate (TDCPP).Chlorinated organophosphate flame retardants (OPFRs) werethe dominant class of flame retardants observed and weredetected in 60 of the 101 samples analyzed. Firemaster 550(FM 550) was detected in 17 samples, as identified by detectionof 2-ethylhexyl-2,3,4,5-tetrabromobenzoate (TBB), bis(2-ethylhexyl)-2,3,4,5-tetrabromophthalate (TBPH), and triphenyl
phosphate (TPP) together in the samples.4 FM 550 also containsseveral isopropylated triaryl phosphate isomers that are tradesecret.14 These isomers were apparent in the GC/MS screeninganalysis but not quantified due to lack of analytical standards.PBDE congeners commonly associated with the PentaBDEmixture were detected in five of the samples examined and werealways found in combination with TPP. Despite the fact thatChemtura ceased production of PentaBDE in 2004, productscontaining this flame retardant are obviously still in active use bythe general public. Four of the five products found to containPBDE congeners were purchased prior to 2004, and the fifthsample was purchased in 2007 from a second-hand store, thusmaking it impossible to determine the original manufacture andpurchase date. Lastly, one sample was found to have significantlevels of TPP but not TBB or TBPH. HPLC-HRMS analysis ofthis sample demonstrated the presence of TPP and threepolybutylated aryl phosphate compounds, which may be fromuse of a flame retardant mixture manufactured by Supresta(Ardsley, NY) and sold commercially as AC073. According toinformation provided in the EPA’s Furniture Flame RetardancyPartnership,15 AC073 consists of TPP (38�48%) and threeproprietary aryl phosphate compounds in concentrations ran-ging from 40 to 46%, 12�18%, and 1�3% for each phosphatecompound. These percentages are very similar to the arearesponses observed for TPP and the butylated aryl phosphatesobserved in our GC/MS and LC/HRMS analyses.Identification of New Flame Retardants. In addition to the
flame retardants described above, we also detected two OPFRs,which to our knowledge, have not been previously identified inthe environmental literature. During our GC/MS analysis of the
Figure 1. Structures of non-PBDE flame retardants detected in polyurethane foam collected from baby products.
foam samples, some samples were found to have either nodetectable levels of the targeted flame retardants or to have verylow levels of TCEP and TCPP. In addition, GC/MS analysis ofsome of these samples revealed chromatographically unresolvedpeaks (i.e., very broad, with significant tailing) eluting afterTCEP and TCPP. We considered it very likely that theseproducts had been treated with some kind of flame retardantsat a significant (percent-by-mass) level in order to meet flameretardancy standards. During the HPLC/HRMS analysis, severalof these samples yielded abundant and chromatographicallyresolved peaks in both positive-ion electrospray and APCImodesfor compounds having mass spectra (e.g., accurate mass andisotope structure) suggestive of a chlorinated organophosphatecompound containing two phosphate groups and six chlorineatoms. Furthermore, it appeared that some samples contained aputative chlorinated organodiphosphate with an [MþH]þ ion at580.91 m/z, while other samples were dominated by a peakgiving an [MþH]þ ion at 636.97m/z. We did not have access toauthentic standards for definitive identification of these com-pounds. However, based on results from both high-resolutionelectrospray ionization and atmospheric pressure chemical ioni-zation, and from MS/MS and MS3 analysis, we propose that onecompound is 2,2-bis(chloromethyl)propane-1,3-diyl-tetrakis(2-chloroethyl)bis(phosphate) (Figure 1). The difference betweenthe predicted (580.9150) and observed (580.9141) m/z forthe [MþH]þ ion of this compound was less than 2 ppm. Thiscompound is known commercially as “V6”. V6 is sold byAlbermarle (Baton Rouge, LA) under the trade name, AntiblazeV6; however, it may also be sold and distributed by other flameretardant companies. A risk assessment conducted by the Eur-opean Commission suggests that V6 is primarily used in auto-mobile foam and has one producer in the European Union.16
According to Albermarle’s material safety data sheet (MSDS) forAntiblaze V6, this mixture contains TCEP as a 10% impurityby weight. V6 is similar in structure to TCEP, containing twobis(2-chloroethyl)phosphate molecules linked by a dichlorodi-methylpropane bridge, which may explain why TCEP is sucha large impurity. We detected the putatively identified V6 in
16 samples, 15 of which also contained significant levels ofTCEP, suggesting that these products may have been treatedwith V6. According to the US EPA’s Inventory Update ReportingDatabase,17 V6 was used in volumes between 1 and 10 millionpounds in reporting years 1990, 1994, and 1998 and between500,000 and 1 million pounds in 2002. V6 was not listed in thedatabase for reporting year 2006, which may indicate that its usein the US has decreased.In addition to V6, the second previously uncharacterized
OPFR compound discovered by HPLC-HRMS in six of thefoam samples appears to be structurally similar to V6 but withpropyl chains connected to the phosphate esters instead of ethylchains. Based on both HPLC/HRMS,MS/MS, andMS3 analysis(Figures S1 and S2 in the Supporting Information), we proposethat this second chemical is 2,2-bis(chloromethyl)propane-1,3-diyl tetrakis(1-chloropropan-2-yl) bis(phosphate). In this manu-script we will refer to this compound as the “U-OPFR”. Asobserved in Figure 2, the difference between the predicted(636.9776) and observed (636.9769) m/z values for monoiso-topic [MþH]þ ions for U-OPFR was less than 2 ppm. We canfind no reference to the use or manufacture of this compound byany chemical company. However, we did find a patent applica-tion submitted by Albermarle in 2008 which describes thepotential application and structure of this chemical.18 Presum-ably the synthesis of this U-OPFR would be very similar to thesynthesis of V6, as these two compounds are structural analogs,suggesting that the U-OPFR would contain TCPP as an im-purity, analogous to the presence of TCEP in V6. In fact, in everysample for which we detected this U-OPFR, we also detectedTCPP.It is also of interest to note that many of the products examined
contained more than one identifiable flame retardant. For exam-ple, in one sample, FM550 andPentaBDEwere detected togetherin appreciable levels, while in another sample both TDCPP andFM 550 were detected. In addition, every sample containingPentaBDE also contained TPP. It appears likely that TPP wasfrequently used in combination with PentaBDE, an observationnot previously reported to our knowledge. Taken together these
Figure 2. Identification of a previously unreported flame retardant, 2,2-bis(chloromethyl)propane-1,3-diyl tetrakis(1-chloropropan-2-yl) bis-(phosphate) “U-OPFR”, and TCPP, in a sample from an infant changing table pad by LTQ-Orbitrap high resolution mass spectrometry. Insetdemonstrates a comparison of the observed and predicted high-resolution mass spectra (MS) for U-OPFR.
observations indicate that some of these flame retardants arebeing used in combinations in commercial products or that thereis contamination in the foam from one batch to the next.Of the 101 products examined in this study, 12 samples were
observed to have significant peaks present in the extracts, butthe identities of the chemicals could not be determined. Andnine samples were observed to have no significant peaks in thechromatograms during the screening step. Therefore, 80% ofthe baby products tested in this study contained a knownand identifiable flame retardant, and all but one of these flameretardants were either brominated or chlorinated.FlameRetardantAssociationswith Products. In general, the
flame retardant chemicals detected were not associated with aparticular type of product, manufacturer, or the year of purchase.An exception to this was the detection of V6 in nursing pillows.We analyzed 11 different samples from nursing pillows, all ofwhich were manufactured by one company. Ten of these samplescontained V6 and were purchased between 2003 and 2008. Theremaining sample was purchased in 2010 and contained primarilyTDCPP as well as appreciable levels of TCPP (1.55 mg/g). Fiveadditional nursing pillows from the same company were pur-chased during the summer of 2010 to determine whether V6 and/or TCEP were present. These samples were screened using GC/MS. The only FR detected was TDCPP, which was found in allfive samples. More information on the flame retardants detectedin each sample can be found in the Supporting Information.Flame Retardant Concentrations in Foam. If authentic
standards were available, we measured the concentrations of thedominant flame retardants detected in the foam samples (Table 1).
TDCPP and PentaBDE were detected in the highest concentra-tions, with average concentrations of 39.2 and 32.3 mg/g, respec-tively (approximately 3�4% by weight). These values are similarto previously reported values of flame retardants in furniture byour group3 but lower than the 32% by weight measurement madeby Hale et al. in polyurethane foam.19 The two brominatedcompounds in the FM 550 formulation were detected at lowerconcentrations than TDCPP and PentaBDE, likely because theyare parts of a mixture. According to the MSDS for FM 550, TBBand TBPH together comprise approximately 50% of the overallmixture. This likely explains why the sum of TBB and TBPH isapproximately 50% of themeasured concentrations of TDCPP andPentaBDE in the foam samples.In general, concentrations of TCEP and TCPP in the samples
were much lower than the concentrations of the other threeprimary flame retardants identified, indicating they may be minorcomponents of flame retardantmixtures, such as V6. In all samplesin which TCEP was detected, V6 or TCPP/TDCPP was alsodetected. In only two samples was TCPP the only identified flameretardant. One sample contained 5.8 mg/g of TCPP, and noother compounds were evident by GC/MS or high resolutionMSanalysis. However, the second sample, which contained onlyTCPP (0.8mg/g), also contained several unidentified chlorinatedcompounds that appeared to be part of a polymeric series, but noconsistent elemental formulas were apparent.XRF Analysis. We investigated whether portable X-ray fluor-
escence (XRF) could be used as a screening tool for predictingthe presence of brominated or chlorinated flame retardantadditives in foam from these products. When both XRF andGC/MS analyses detected bromine in the foam samples, asignificant correlation (p < 0.001) was observed (Figure 3a). Insamples containing FM550, XRF-measured bromine generallyoverpredicted the GC/MS-measured bromine by about 100%.This overprediction is consistent with that found earlier by Allenet al. 12 and may be due to differences in the sample matrix as thecalibration standards used with the XRF device are hard plastics.However, there were seven samples in which XRF analysesdetected bromine ranging from 1.4�3.4% by weight, but GC/MSdetected only chlorinated OPFRs. This suggests that there areeither some instances in which false positives are generated forbromine in polyurethane foam by XRF, possibly due to inter-ferences by other elements, or there are unknown brominatedcompounds present in some of these foam samples that were notaccounted for by GC/MS analysis.As seen in Figure 3b, there was no significant relationship
observed between XRF- and GC/MS-measured chlorine in thesesamples. The fact that we detected V6, and the U-OPFR, butcould not quantify themwithout an authentic standard, was likelya contributing factor for the poor relationship between the XRFand GC/MS analyses. While removing these compounds fromthe correlation analysis resulted in a higher correlation coeffi-cient, the slope was still not significant (data not shown). Also, inthree samples XRF-measured chlorine ranged from 1.2�3.3% byweight, yet GC/MS determined that only BFRs were present.Chlorinated impurities present in toluene diisocyanate (TDI), astarting material for the synthesis of polyurethane foam, may beresponsible for these chlorine signals and would not have beendetectable in the GC/MS analysis. These TDI impurities mayalso have contributed to the much higher concentrations of XRF-measured chlorine observed (2.2�23.7%) compared to theGC/MS results for the OPFRs. Based on these results, we believethat XRF is generally a useful screening tool for identifying the
Figure 3. Correlation between GC/MS and XRF measured bromine(A) and chlorine (B).
presence of BFRs in foam; however, additional work is needed tounderstand the extent of its use as an effective screening tool forchlorinated flame retardants.Infant’s Exposure Potential and Health Concerns. This
study found that more than 80% of the baby products testedcontained a halogenated flame retardant additive, many of whichwere chlorinated OPFRs. This suggests these products could besources of flame retardant exposures in indoor environments,particularly to infants that come in close contact with theseproducts. In 2006, the Consumer Product Safety Commission(CPSC) released a Risk Assessment of Flame Retardant Chemi-cals in Upholstered Furniture Foam, which included TDCPP.20
This CPSC report states that “...upholstered furniture manufac-tured with TDCPP treated foam might present a hazard toconsumers, based on both cancer and non-cancer end points”.The CPSC estimate of children’s exposure to TDCPP fromtreated furniture was five times higher than the agency’sacceptable daily intake (i.e., the Hazard Index was 5). Almost99% of this exposure was from inhalation of TDCPP volatilizedfrom treated furniture (air concentrations were predicted nearfurniture and in rooms rather than measured, a major source ofuncertainty). TDCPP was the most common flame retardantidentified in this screening study, with concentrations verysimilar to those reported in upholstered furniture.3 For severalreasons, infants exposure to TDCPP could be higher than theexposure calculated by the CPSC. Infants have smaller bodymasses relative to the average child or adult used in theirassessment. Infants spend a greater proportion of their time inintimate contact with these materials (e.g., infant sleep posi-tioners, car seats, nursing pillows) over a longer daily time periodthan the 3 h assumed in the CPSC report. In addition, newstudies are suggesting that exposure to semi-volatile organiccompounds may be occurring from equilibrium partitioningbetween the indoor gas phase and skin surfaces/clothing, whichcan lead to accumulation via skin absorption.21 TDCPP has beenshown to be efficiently absorbed through the skin of rodents,with as much as 85% of the dose absorbed dermally.22 Therefore,exposure of infants to TDCPP, and likely other flame retardants,may be greater than the Hazard Index of 5 calculated by theCPSC. Further research is warranted to investigate infant ex-posure to flame retardants in these products, particularly sinceinfants are in a very sensitive development stage andmay bemoresusceptible to adverse effects than an older child or adult.Previous studies have shown that TDCPP, and its brominated
analogue tris (2,3-dibromopropyl) phosphate, were previouslyused as flame retardants in children’s sleepwear. However, thisuse was discontinued after studies found that children wearingthese clothes absorbed TDBPP.23 Both TDBPP and TDCPPwere observed to be mutagenic in the Ames assay, particularlyafter metabolism.24 Rats exposed to TDCPP were found to haveincreased incidences of tumors,25 and a recent study also foundthat TDCPP was as potent a neurotoxicant as chlorpyrifos usingan in vitro assay.26 One study found that TDCPP levels in housedust were significantly correlated with reduced thyroid hormonelevels and increased levels of prolactin in men.27 And one studydetected TDCPP and several other OPFRs at concentrationssimilar to PBDEs in US house dust,3 suggesting chronic exposureto the population is occurring on a daily basis. In addition, theEuropean Chemical Bureau of the European Union considersTCEP to be a category 3 carcinogen.28
This study adds to our understanding of flame retardantsin consumer products. The comparison of XRF and GC/MS
measurements for bromine confirm previous results that thistechnology is generally useful for screening brominated flameretardants in polyurethane foam. The results for chlorine havenot been previously reported and suggest that additional researchis needed before XRF can reliably screen for chlorinated flameretardants in polyurethane foam. Levels of up to 12.5% ofTDCPP were found in one product, while other products werefound to contain up to three different retardants in one product.Lastly, we have identified two flame retardants previouslyunreported in the environment. Further studies are also war-ranted to determine whether V6 and the U-OPFR are present inindoor environments and whether human exposure is a concern.
’ASSOCIATED CONTENT
bS Supporting Information. High resolution tandem massspectra and proposed fragmentation mechanisms and pathwaysrelevant to the identification of the putative U-OPFR compounddescribed in the manuscript are available in the SupportingInformation. We also include a table summarizing the types andrelative abundances of flame retardant chemicals analyzed in allsamples measured in the present study. This material is availablefree of charge via the Internet at http://pubs.acs.org.
The XRF analyzer was provided for this study at no cost byJack Hanson (Olympus Innov-X Systems, and HMC AnalyticalInstrumentation, Livermore, CA). The authors would like tothank Ms. Courtney Walker for her assistance in the XRFanalysis. Dr. Heather M. Stapleton was funded and supportedby a grant from the National Institute of Environmental HealthSciences, R01ES016099. Drs. Ferguson and Stapleton were alsopartially supported by a donation from Fred and Alice Stanback.Dr. Webster is partly supported by R01ES015829 from theNational Institute of Environmental Health Sciences.
’REFERENCES
(1) Tullo, A. Great Lakes to phase out flame retardants. Chem. Eng.News 2003, 81 (45), 13–13.
(2) EU, Directive 2003/11/EC of the European Parliament and ofthe Council. In Official Journal of the European Union: 2001; Vol.Directive 76/769/EEC, pp 45-46.
(3) Stapleton, H.M.; Klosterhaus, S.; Eagle, S.; Fuh, J.; Meeker, J. D.;Blum, A.; Webster, T. F. Detection of Organophosphate Flame Re-tardants in Furniture Foam and US House Dust. Environ. Sci. Technol.2009, 43 (19), 7490–7495.
(4) Stapleton, H. M.; Allen, J. G.; Kelly, S. M.; Konstantinov, A.;Klosterhaus, S.; Watkins, D.; McClean, M. D.; Webster, T. F. Alternateand new brominated flame retardants detected in US house dust.Environ. Sci. Technol. 2008, 42 (18), 6910–6916.
(5) Requirements, test procedure and apparatus for testing the flameretardance of resilient filling materials used in upholstered furniture. InCaliforniaDepartment of Consumer Affairs: NorthHighlands, CA, 2000.
(6) Weil, E. D.; Levchik, S. V. Commercial flame retardancy ofpolyurethanes. J. Fire Sci. 2004, 22 (3), 183–210.
(7) Hites, R. A. Polybrominated diphenyl ethers in the environmentand in people: A meta-analysis of concentrations. Environ. Sci. Technol.2004, 38 (4), 945–956.
(8) Hale, R. C.; Alaee, M.; Manchester-Neesvig, J. B.; Stapleton,H. M.; Ikonomou, M. G. Polybrominated diphenyl ether flame retar-dants in the North American environment. Environ. Int. 2003, 29 (6),771–779.(9) Chen, S. J.; Ma, Y. J.; Wang, J.; Chen, D.; Luo, X. J.; Mai, B. X.
Brominated Flame Retardants in Children’s Toys: Concentration,Composition, and Children’s Exposure and Risk Assessment. Environ.Sci. Technol. 2009, 43 (11), 4200–4206.(10) Lorber, M. Exposure of Americans to polybrominated diphenyl
Harvey, E.; La Guardia, M.; McClean, M. D.; Webster, T. F. Humanexposure to PBDEs: Associations of PBDE body burdens with foodconsumption and house dust concentrations. Environ. Sci. Technol. 2007,41 (5), 1584–1589.(12) Allen, J. G.; McClean, M. D.; Stapleton, H. M.; Webstert, T. F.
Linking PBDEs in house dust to consumer products using X-rayfluorescence. Environ. Sci. Technol. 2008, 42 (11), 4222–4228.(13) Harrad, S.; de Wit, C. A.; Abdallah, M. A. E.; Bergh, C.;
Bjorklund, J. A.; Covaci, A.; Darnerud, P. O.; De Boer, J.; Diamond,M. L.; Huber, S.; Leonards, P.; Mandalakis, M.; Ostman, C.; Haug, L. S.;Thomsen, C.; Webster, T. F. Indoor Contamination with Hexabromo-cyclododecanes, Polybrominated Diphenyl Ethers, and PerfluoralkylCompounds: An Important Exposure Pathway for People? Environ. Sci.Technol. 2010, 44 (9), 3221–3231.(14) Chemtura.Material Safety Data Sheet for Firemaster 550; 2006;
Profiles of Chemical Flame Retardant Alternatives. In EPA 742-R-05-002A: 2005.(16) EC, Risk Assessment Report on 2,2-bis(chloromethyl)trimethylene
bis[bis(2-chloroethyl)phosphate] (V6). European Comission: Brussels,2007.(17) http://www.epa.gov/iur/.(18) Mack, A. G.; Chai, Z. Pat. Appl. WO 208/100316 Al,
Phosphate-Containing Flame Retardants, 2008.(19) Hale, R. C.; La Guardia, M. J.; Harvey, E.; Mainor, T. M.
Potential role of fire retardant-treated polyurethane foam as a source ofbrominated diphenyl ethers to the US environment. Chemosphere 2002,46 (5), 729–735.(20) Babich, M. A. CPSC Staff Preliminary Risk Assessment of
Flame Retardant (FR) Chemicals in Upholstered Furniture Foam. U.S.Consumer Product Safety Commission, 2006, pp 1�129.(21) Weschler, C. J.; Nazaroff, W. W. Semivolatile organic com-
pounds in indoor environments. Atmos. Environ. 2008, 42 (40), 9018–9040.(22) Hughes, M. F.; Edwards, B. C.; Mitchell, C. T.; Bhooshan, B. In
vitro dermal absorption of flame retardant chemicals. Food Chem.Toxicol. 2001, 39 (12), 1263–1270.(23) Blum, A.; Gold, M. D.; Ames, B. N.; Kenyon, C.; Jones, F. R.;
Hett, E. A.; Dougherty, R. C.; Horning, E. C.; Dzidic, I.; Carroll, D. I.;Stillwell, R. N.; Thenot, J. P. Children Absorb Tris-Bp Flame-Retardantfrom Sleepwear - Urine Contains Mutagenic Metabolite, 2,3-Dibromo-propanol. Science 1978, 201 (4360), 1020–1023.(24) Gold, M. D.; Blum, A.; Ames, B. N. Another Flame-Retardant,
Tris-(1,3-Dichloro-2-Propyl)-Phosphate, and Its Expected MetabolitesAre Mutagens. Science 1978, 200 (4343), 785–787.(25) NRC, Toxicological Risks of Selected Flame-Retardant Chemicals.
National Academy Press: Washington DC, 2000.(26) Dishaw, L. V.; Powers, C.M.; Ryde, I. T.; Roberts, S. C.; Seidler,
F. J.; Slotkin, T. A.; Stapleton, H.M., Is the PentaBDEReplacement, Tris(1,3-dichloro-2-propyl) Phosphate (TDCPP), a Developmental Neu-rotoxicant? Studies in PC12 Cells. Toxicol. Appl. Pharmacol. 2011, doi:10.1016/j.taap.2011.02.005.(27) Meeker, J. D.; Stapleton, H. M. House Dust Concentrations of
Organophosphate Flame Retardants in Relation to Hormone Levelsand Semen Quality Parameters. Environ. Health Perspect. 2010, 118 (3),318–323.
(28) Tris (2-chloroethyl)phosphate, TCEP, Summary Risk AssessmentReport; 793/93, C. R. E. N.; Federal Institute for Occupational Safetyand Health: Dormund, Germany, 2008.
